March 2005 - Darwin-dev - lists.apple.com

Re: pthread_join and pthread_detach
by Nicolas.Castagne＠imag.fr 23 Mar '05

23 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Importance: Normal User-agent: SquirrelMail/1.4.2 >From Paul Forgey : > What you are doing sounds frightening. Yeah, I knew ;-) Anyhow, I have arranged my program in a more reliable way yesterday - something like what you kindly proposed, thx. In fact, I was mainly wondering if I was encoutering a bug on the Mac pthread lib. And I still think there may be one, or at least a lake in the man pages. I enclose an example, as requested by matt. When I select a small amount of time work, that is when the main thread 'joins' the thread1 before the thread 1 calls pthread_detach, the program crashes. Well, not on the first trial, but on the second or the third. In a few words, it seems that : "You cannot call pthread_detach on a thread while a call to pthread_join is pending" This, at least, should be documented, no ? All the best Nicolas _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com main.cpp

1 0

Re: AppleRTL8139Ethernet.kext
by Finlay Dobbie 23 Mar '05

23 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:references; b=cEfSlLULcBhiTzAHy2nUrhjywXdsM9hhk3MFeU8k2eHcT/215atuAe1lPEpk+NOO+4fHOynfP10wXkrR2NHqYf9m29nRyJn0ty49RPx47LB9gOb8cQYnji8DC/TJORsvMAO/txNKU2CRDVM+eSZCitnXE/Dcp6jwcaTILnmE/3E= On Tue, 22 Mar 2005 13:25:33 -0800, Walt Pawley <walt(a)wump.org> wrote: > I had hoped to find this driver code in the Darwin source. It doesn't seem > to be there. ??? Not all drivers are open-source. Some contain Apple IP that they deem to be confidential, or third party IP that they can't distribute in source form. -- Finlay _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: pthread_join and pthread_detach
by Matt Watson 23 Mar '05

23 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Mar 22, 2005, at 9:30 AM, Nicolas.Castagne(a)imag.fr wrote: Thanks, matt. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… If it is, well... It is not documented in the man pages... Does it correspond with the POSIX standard, or should I register it as a bug ? It's hard to say without a concrete example. Can you post a self- contained program (with a main()) that exhibits this problem? This email sent to site_archiver(a)lists.apple.com

1 0

Re: Seems weird behaviors of mkfifo
by Tian-Jian Barabbas Jiang 23 Mar '05

23 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

AppleRTL8139Ethernet.kext (correction)
by Walt Pawley 22 Mar '05

22 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I originally wrote, 'The kernel then starts logging messages proclaiming "Weird Interrupt..."' This is not true. The actual text is "Weired Interrupt". Out of curiousity, anyone know from where that bit arises? -- Walter M. Pawley <walt(a)wump.org> Wump Research & Company 676 River Bend Road, Roseburg, OR 97470 541-672-8975 _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

AppleRTL8139Ethernet.kext
by Walt Pawley 22 Mar '05

22 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I have a beige G3 I'd like to run a RealTek 8139 Ethernet PCI adapter in (since I have a couple of them). There's a driver for Jaguar (and earlier) downloadable from RealTek ... and it works ... sort of. Curiously, it will trundle along for quite some time, having no apparent difficulty until one attempts to move something like an image file of a full CD, say, through it. The kernel then starts logging messages proclaiming "Weird Interrupt..." and it subsequently becomes very dodgie. Both cards seem to behave the same. I originally thought the first one was broken but it's had no trouble functioning in a FreeBSD box for a number of moons now. I have been led to believe that the 8139 was workable without a third party driver in Panther. That got me interested in looking at the list of extensions on my G5, where I bumped into AppleRTL8139Ethernet.kext, which probably explains why the cards are supposed to work in Panther. I had hoped to find this driver code in the Darwin source. It doesn't seem to be there. ??? Also, does anyone happen to know whether AppleRTL8139Ethernet.kext will function under Jaguar? That, and whether it actually works? -- Walter M. Pawley <walt(a)wump.org> Wump Research & Company 676 River Bend Road, Roseburg, OR 97470 541-672-8975 _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Seems weird behaviors of mkfifo
by Paul Forgey 22 Mar '05

22 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Mar 22, 2005, at 10:32 AM, Tian-Jian Barabbas Jiang wrote: Hi all, mkfifo -m 0666 /tmp/pipe tail -f /tmp/pipe and then cat > /tmp/pipe > a > b > c cat > /tmp/pipe > d > e > f this time the 'tail -f' will show 'd', 'e', 'f' one by one immediately. Thank you for your patience to read this mail. :) Best Regards, Tian-Jian Jiang * Server.cpp #include <iostream> #include <fstream> #include <string> // for umask() and mkfifo() #include <sys/types.h> #include <sys/stat.h> using namespace std; int main() { umask(0); int returnValue = mkfifo("/tmp/pipe", 0666); previousString = outputString; in.close(); } } else cout << "mkfifo failed." << endl; return 0; } * Client.cpp #include <iostream> #include <fstream> #include <string> using namespace std; int main() { string inputString(""); ofstream out("/tmp/pipe"); while(cin >> inputString) { out << inputString << endl; if(inputString == "/quit") break; } out.close(); return 0; } _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/paulf%40aphrodite.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… The behavior you are seeing is normal for any posix system. As far as your shell example goes, that's the effects of buffering. The reader won't read until a buffer is written with an underlying write () call to the pipe. When that happens is entirely dependent on the shell. The SIGPIPE's in your C code are normal. It means you tried to read from the read end of a pipe after the writer closed it. You can set SIGPIPE to SIG_IGN and have the read() operation fail with EPIPE instead. Note what you are trying to do will only work with one client at a time, and unless you arrange for some sort of handshaking, nothing will stop more than one client from trying and garbling your input. Normally two pipes are set up to handshake both ways between the client and the server. I've just wrote two small programs of client/server mechanism based on mkfifo, and I've encountered two strange situations. 1. The named pipe was unstable for my programs. I've experienced "Broken pipe." frequently while sending messages from the client. 2. Reading pipe was not that atomic. For example, If I just test by commands rather than handcraft programs, for the first try in another screen, 'tail -f' won't show anything until I leave 'cat'. But if I give it a try again, For my programs, if I want the server to read and to print the client's messages immediately, I have to repeat opening the file of named pipe. For more detail, please refer to comments in my programs below. And FYI, I also have tested these programs on FreeBSD 4.10 and FreeBSD 5.3, they have worked fine. I'm wondering if these are Darwin's problems, so I post this mail here. if(returnValue == 0) { string previousString(""); while(previousString != "/quit") { // It's easy to see the strange behavior in the following codes. I supposed the normal usage is to open file only once outside of the while-loop. // Unfortunately, I have to open the file again and again to make this program work on Darwin. But I don't need to do this on FreeBSD. ifstream in("/tmp/pipe"); string outputString(""); in >> outputString; if(outputString.length() > 0 && outputString != previousString) cout << "> " + outputString << endl; This email sent to paulf(a)aphrodite.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: pthread_join and pthread_detach
by Paul Forgey 22 Mar '05

22 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com struct MyThread { pthread_t thread; pthread_mutex_t mutex; int refCount; }; On Mar 22, 2005, at 9:30 AM, Nicolas.Castagne(a)imag.fr wrote: Hi all, I set up a solution on Linux that ran properly. Here is an example of the code : // THREAD 1 routine // (thread 1 pthread_t is thread1_id) void thread1_routine(void *data) { // do some work // wait for thread 2 // As an example, thread2 is here always joined. // However, in my app, it is joined only in some cases int res = pthread_join(thread2_id, NULL); // THREAD 2 routine // (thread 2 pthread_t is thread2_id) void thread2_routine(void *data) { // do some work // before terminating, detach the thread to be sure its storage is reclaimed int res = pthread_detach(thread2_id, NULL); if( res != 0) { switch(res) { case ESRCH : printf("the thread is not valid \n"); break; case EINVAL : printf("the thread is already detached\n); break; } pthread_exit(NULL); } As far as I chercked, a code like this one does not run on mac os X. The call to pthread_join results in : mach_port_mod_refs(reply_port) failed: (os/kern) invalid name mach_port_deallocate(kernel_thread) failed: (os/kern) invalid name (then, the program aborts). Thx, Nicolas _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/paulf%40aphrodite.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… What you are doing sounds frightening. Even if that did work, there's no guarantee that pthread_join or pthread_detach ever reasonably expect to work atomically if multiple threads potentially call it at the same time on the same thread. Why do you need to do this? You can call pthread_detach or pthread_join on an already exited thread, and if you know ahead of time you will never need to join the thread, just create it detached in the first place. On the other hand, does your model really need to retain and release reference counts to the thread? From the code you've quoted, no. Just join on the thread from the first thread and be done with it. But if there's more to it than that, consider something like this: Wrap your thread handle around a structure containing a mutex and ref count, and write functions to create, retain and release the thread appropriately. in my app, some threads that cannot be detached a-priori : sometimes, they have to be joinable, but sometimes they can be detached (no other thread will wait for them). At the same time, I needed to be sure that storage of all the thread are reclaimed when they terminate. On linux, I always detached the thread that may be joined JUST BEFORE it exists. In case this thread was not awaited (no pending call to pthread_join() by another thread), its storage was reclaimed when terminating. In case this thread was already awaited (joined) by another thread, the pthread_join in the other thread resulted in a error EINVAL indicating that the thread was detached. However, the storage was still reclaimed when terminating, and the pthread_join exited when the joined thread terminates (or just before). In a few words, it was 'OK'. // in case the thread2 was not already terminated, EINVAL is returned... // in case the thread2 wa already terminated, ESRCH is returned... if( res != 0) { switch(res) { case ESRCH : printf("pthread_join error the thread is not valid\n"); break; case EINVAL : printf("the thread is already detached\n"); break; case EDEADLK : printf("called by the calling thread\n"); break; default : couterr <<"unknown"<<endl; break; } } } and sometimes in : mach_port_deallocate(kernel_thread) failed: (os/kern) invalid name *** malloc[1243]: Deallocation of a pointer not malloced: 0x580bc00; This could be a double free(), or free() called with the middle of an allocated block; Try setting environment variable MallocHelp to see tools to help debug Indeed, it seems that we cannot call pthread_detach on a thread while this thread is already by another thread awaited within a pthread_join. Is that true ? If it is, well... It is not documented in the man pages... Does it correspond with the POSIX standard, or should I register it as a bug ? This email sent to paulf(a)aphrodite.com This email sent to site_archiver(a)lists.apple.com

1 0

Seems weird behaviors of mkfifo
by Tian-Jian Barabbas Jiang 22 Mar '05

22 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

pthread_join and pthread_detach
by Nicolas.Castagne＠imag.fr 22 Mar '05

22 Mar '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Importance: Normal User-agent: SquirrelMail/1.4.2 Hi all, in my app, some threads that cannot be detached a-priori : sometimes, they have to be joinable, but sometimes they can be detached (no other thread will wait for them). At the same time, I needed to be sure that storage of all the thread are reclaimed when they terminate. I set up a solution on Linux that ran properly. On linux, I always detached the thread that may be joined JUST BEFORE it exists. In case this thread was not awaited (no pending call to pthread_join() by another thread), its storage was reclaimed when terminating. In case this thread was already awaited (joined) by another thread, the pthread_join in the other thread resulted in a error EINVAL indicating that the thread was detached. However, the storage was still reclaimed when terminating, and the pthread_join exited when the joined thread terminates (or just before). In a few words, it was 'OK'. Here is an example of the code : // THREAD 1 routine // (thread 1 pthread_t is thread1_id) void thread1_routine(void *data) { // do some work // wait for thread 2 // As an example, thread2 is here always joined. // However, in my app, it is joined only in some cases int res = pthread_join(thread2_id, NULL); // in case the thread2 was not already terminated, EINVAL is returned... // in case the thread2 wa already terminated, ESRCH is returned... if( res != 0) { switch(res) { case ESRCH : printf("pthread_join error the thread is not valid\n"); break; case EINVAL : printf("the thread is already detached\n"); break; case EDEADLK : printf("called by the calling thread\n"); break; default : couterr <<"unknown"<<endl; break; } } } // THREAD 2 routine // (thread 2 pthread_t is thread2_id) void thread2_routine(void *data) { // do some work // before terminating, detach the thread to be sure its storage is reclaimed int res = pthread_detach(thread2_id, NULL); if( res != 0) { switch(res) { case ESRCH : printf("the thread is not valid \n"); break; case EINVAL : printf("the thread is already detached\n); break; } pthread_exit(NULL); } As far as I chercked, a code like this one does not run on mac os X. The call to pthread_join results in : mach_port_mod_refs(reply_port) failed: (os/kern) invalid name mach_port_deallocate(kernel_thread) failed: (os/kern) invalid name and sometimes in : mach_port_deallocate(kernel_thread) failed: (os/kern) invalid name *** malloc[1243]: Deallocation of a pointer not malloced: 0x580bc00; This could be a double free(), or free() called with the middle of an allocated block; Try setting environment variable MallocHelp to see tools to help debug (then, the program aborts). Indeed, it seems that we cannot call pthread_detach on a thread while this thread is already by another thread awaited within a pthread_join. Is that true ? If it is, well... It is not documented in the man pages... Does it correspond with the POSIX standard, or should I register it as a bug ? Thx, Nicolas _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0